Keywords: simulation, Matlab, virtual prototyping, highway intersections, traffic collisions, intersection collision warning systems.

A major challenge that faces the design of any new intersection collision warning system is the high cost and risk involved with producing and testing a prototype for the proposed system. Furthermore, there is tradeoff between the field performance of the system and the specifications selected for its components; and therefore, it is necessary to have more flexibility in testing the system under different hardware and software specifications. This paper provides details on simulating an intersection collision warning system that utilizes a pair of radar sensors to detect vehicles approaching an intersection in addition to a real-time algorithm that determines their speeds, acceleration rates and the rate-of-change in acceleration in order to estimate the time required for each approaching vehicle to reach the intersection and compare it with the time required for the equipped vehicle to clear the intersection. By comparing the two times together, the system displays a message for the driver of the equipped vehicle when departure is safe.

The Matlab simulation codes, that were created, were used to select the minimum range resolution and azimuth angle resolution for the radar sensors utilized by the system, where it was found that for a typical speed of 60 km/h, selecting a range precision of 5 cm (or better) and azimuth angle resolution of 0.1° (or better) will ensure acceptable performance of the proposed intersection collision warning system without any significant error.

The Matlab simulation codes were also used to select the acceleration models that should be used by the system's algorithm for both the approaching and the equipped vehicles, and it was found that for an approaching vehicle, the algorithm should assume a variable acceleration rate with a constant rate of change in acceleration. As for the equipped vehicle, it was found that the linear decay acceleration model, which assumes that the acceleration rate decreases with the increase of the speed of the vehicle, is more suitable and gives more realistic results without significant errors.

The simulation findings for the range and azimuth angle resolutions were found to be consistent with previous research, which validates the simulation codes and gives potential for further research where those codes may be utilized as a less expensive and safer alternative for an actual prototype, especially in the early stages of building a proposed intersection collision warning system where different specifications should be selected such as the opening angle, maximum range, update rate and data latency.

purchase the full-text of this paper (price £20)

go to the previous paper
go to the next paper
return to the table of contents
return to the book description
purchase this book (price £125 +P&P)